PROMPT GUIDE
How to Write AI Video Prompts (Motion, Camera and Time)
Unlike image prompts, AI video prompts describe the passage of time. Rather than what is in the frame, they tell the AI what moves and how, how the camera moves, and what mood the shot ends in. This guide explains the subject, motion, camera, mood structure that works with almost any AI video generator, with examples.
Last updated: approx. 3 min read
The short answer
Write AI video prompts as two to four short English sentences in the order subject, motion, camera, mood. Use one motion per shot and add speed words such as slowly or gently, so the video moves as intended without breaking.
How are video prompts different from image prompts?
An image prompt describes what is inside a single picture. A video prompt adds motion, camera and time. If you only list appearance, the AI does not know what to move, and you get an almost frozen video or one where everything moves randomly.
AI video generators also tend to understand short English sentences better than the tag lists common in image generation (1girl, smile, long hair and so on). For text to video, describe the look too; for image to video, write only the motion and camera.
For when to use tags and when to use sentences: Tag Prompts vs Natural Language Prompts (Which to Use?) →
What order should a video prompt follow? (Subject, motion, camera, mood)
Writing a video prompt like one line of a film script keeps it organized: first who or what, then how it moves, then how the camera moves, and finally the light and atmosphere.
A girl with a red ribbon walks slowly along a riverside path. The camera tracks alongside her at eye level. Warm sunset light, a gentle breeze moves her hair and skirt.What it does: the girl walks slowly by the river while the camera moves beside her at eye level, in warm sunset light with a breeze moving her hair and skirt.
Steps for writing a video prompt
- Be able to describe the video in one sentence (for example, following a girl walking along a river at dusk from the side)
- Write the subject briefly. For image to video, just refer to the person in the picture as "the girl"
- Write one motion, such as walking, turning, looking up or waving, and add the speed (slowly or quickly)
- Write the camera move: static camera, push in, tracking and so on
- Add light, weather and atmosphere last. Make a few versions, and soften the wording if the motion is too big
Why does packing in motions break the video? (One motion per shot)
A common failure is cramming too many motions into one video. If you write "turns around, smiles, waves and starts running", the AI tries to fit everything into a few seconds, and the body twists or the face changes. Keep one motion per clip, or at most a main motion plus a small one (hair swaying, blinking), and make the rest as separate clips to join later.
How do you describe speed and strength of motion?
Adverbs change the size of the motion a lot. If the video breaks, first make the motion smaller with slowly or gently.
Example prompts by scene
Static camera. The girl sits by the window reading a book. She slowly turns a page and glances outside as raindrops run down the glass. Soft gray daylight.What it does: a still camera shows the girl reading by the window; she turns a page and glances outside while raindrops run down the glass, in soft overcast light.
The camera slowly pulls back from a close-up of the girl to reveal a wide field of sunflowers under a bright summer sky. Her hair moves gently in the wind.What it does: the camera moves back from a close-up of the girl until a wide sunflower field under a summer sky appears, while her hair moves gently in the wind.
Rinka's tips
- Negative wording (no..., does not...) often works poorly. Describe the state you want instead, such as "static camera"
- On services with a negative prompt field, adding breakage words such as blurry, distorted face and extra fingers may reduce problems
- On services that support dialogue or sound, put the line in quotation marks and say who speaks it
For a list of camera move terms: Camera Movement Prompts for AI Video Generators →
For the steps to turn an image into a video: How to Use Image-to-Video AI (Steps to Animate a Still Image) →
For fixes when things go wrong: Why AI Videos Break and How to Fix Them →
FAQ
Frequently Asked Questions
Should AI video prompts be written in English?
More services now support other languages, but motion and camera instructions are often understood better in English. Short, simple English sentences are enough.
How long should a video prompt be?
About two to four sentences. One sentence each for subject, motion, camera and mood is enough; making it too long tends to create conflicting instructions.
Can I use image generation tags such as 1girl?
You can, but AI video generators tend to understand motion better in sentences. Use tags only as a small supplement for the look, and write the motion as sentences.
The same prompt gives a different video every time.
AI video generators produce different results each time, even with the same instructions. Plan to make several and choose, and save the prompt when you get motion you like.
RINKA WORLD
Play in Rinka World
MORE
